|
Impact of cancer-related decision aids |
O'Brien M A, Villias-Keever M, Robinson P, Skye A, Gafni A, Brouwers M, Charles C, Baldassarre F, Gauld M |
|
|
Authors' objectives The objectives of the review were to describe the different cancer-related decisions aids that have been developed, and to evaluate the effectiveness of these interventions.
Searching The following electronic databases were searched: MEDLINE (from 1977 to April 2001); HealthSTAR, Cancerlit, CINAHL, Sociological Abstracts, PsycINFO (from 1977 to August 2000); EMBASE (from 1995 to August 2000) and the Cochrane Library (Issue 3, 2000). The search terms used were provided. The reference lists of the included studies and relevant published reviews were also checked, as were the personal files of experts in the field. There were no languages restrictions. Unpublished studies and those published only in abstract form were excluded.
Study selection Study designs of evaluations included in the reviewStudies of any design were eligible for inclusion.
Specific interventions included in the reviewFor the purposes of this review a decision aid was defined as an intervention designed primarily to help patients, or patients and clinicians together, with making cancer-related health care decisions when options are available for prevention, screening and treatment. At a minimum, it should target some component of decision-making (e.g. information exchange, involvement in the decision process). In addition, the intervention had to be above and beyond usual care and the patient must have been involved in the use of the instrument. Decision aids aimed solely at the clinician, e.g clinical guideline algorithms, were not to be included in the review. The effectiveness studies included in the review evaluated brochures, audiotapes, videotapes, interactive computer programs, educational scripts, decision boards, counselling and informal decision analysis.
Participants included in the reviewPatients suffering from cancer, or cancer patients and their physicians, were included in the review. Patients using decision aids for hormone replacement therapy, benign prostatic hyperplasia, or smoking cessation were specifically excluded from the review. Breast cancer and prostate cancer were the most frequently studied cancers in the review, accounting for 33 of the 39 studies. All the included studies were of adult populations.
Outcomes assessed in the reviewSpecific outcomes were not predefined for the review. Across the included studies, the patients decisions, knowledge, anxiety, depression, satisfaction and acceptability of the decision aid were the most frequent outcome measures.
How were decisions on the relevance of primary studies made?Seven trained reviewers were involved in an initial screening process, which resulted in 10% of all citations being screened independently by two reviewers. Two reviewers independently assessed full-text articles. For each full English-language article retrieved, two reviewers made independent decisions or whether to include or exclude it. Any disagreements were resolved with the aid of a third reviewer. Non-English articles were selected by a single reviewer proficient in that language.
Assessment of study quality The quality assessment of the randomised controlled trials (RCTs) was based on three scales: the Jadad scale, the Downs and Black scale, and the Guyatt scale. For other study designs (non-randomised controlled trials, non-concurrent cohort studies and case series), the Downs and Black scale only was used. Two independent reviewers performed the quality assessment.
Data extraction The members of the review team developed, pilot tested and revised the data extraction forms. Two independent reviewers extracted the data and resolved any discrepancies by consensus. Data were extracted for each relevant outcome as reported in each individual study.
Methods of synthesis How were the studies combined?Due to a significant degree of clinical heterogeneity, the studies were not pooled in a meta-analysis but discussed narratively. Case series and studies of a pre-test post-test design were grouped together.
How were differences between studies investigated?The results of the studies were discussed according to the type of intervention, subgrouped by study design and within each subgroup by outcome measure.
Results of the review Thirty-nine studies evaluating the effectiveness of decision aids were included in the review. Of these, 16 were RCTs, 4 were non-randomised controlled trials, 2 were nonconcurrent cohort studies, 6 were of a pre-test post-test design and 11 were case series.
The three scales assessing the quality of the RCTs yielded similar results. They revealed common areas of weakness in the included trials: in particular, all trials scored poorly for internal validity.
The following interventions were evaluated in RCTs: decision aid brochures, educational scripts, audiotapes, videotapes, interactive computer programs, counselling, informal decision analysis, decision boards and complex decision aids (several different components).
Overall, among RCTs, the decision aids appeared to increase knowledge and patient involvement in decision-making. Anxiety and depression scores did not appear to be increased. Other results from studies that included a concurrent comparison were that decision aids decreased decisional conflict or uncertainty, and had an influence on decision-making. In patients making prostate cancer screening decisions, significantly fewer men decided to proceed with screening after receiving a decision aid. For decisions about screening, there was insufficient evidence to indicate which type of decision aid might be more effective.
Authors' conclusions Decision aids appeared to be helpful without increasing anxiety or depression, particularly in the context of decisions about screening, but there was a lack of evidence relating to the effectiveness of decision aids for decisions related to cancer treatment.
CRD commentary This review addressed a very broad and loosely defined group of interventions. Consequently, the populations studied, the study designs eligible and the outcomes measures reviewed were not defined narrowly a priori. However, this was not inappropriate. The literature search appeared to be comprehensive in terms of published literature, with a wide range of databases being searched without any language restrictions imposed. The exclusion of unpublished studies and those published only in abstract form may have subjected the review to publication bias. Several methods were used to minimise reviewer bias in the conduct of the review. The validity assessments were thorough and the results were described in some detail. However, the results of these assessments were not integrated fully into the findings of the review, thus limiting their usefulness.
Details of the primary studies were tabulated within the review. The synthesis allowed the reader to access the results from the RCTs separately from the other study designs, yet still provided information on common outcomes from all study designs. This is useful given the broad nature of the interventions and outcomes measures involved in this therapeutic area. Overall, the authors' general conclusions are supported by the review. However, given the disparate nature of the interventions, studies and outcomes measures, it is important to read the reviews' findings in detail for a meaningful understanding of them.
Implications of the review for practice and research Practice: The authors stated that decision aids are helpful for some cancer screening decisions.
Research: The authors stated that the early stage of development of this field, along with the gaps in our knowledge determined by this systematic review, underline the need for further research. Areas specified were identifying when decision-making occurs and who is involved; identifying the key features of decision-making; and the use of these to develop effective interventions and to identify pertinent outcomes measures.
Funding Agency for Healthcare Research and Quality, contract number 290-97-0017.
Bibliographic details O'Brien M A, Villias-Keever M, Robinson P, Skye A, Gafni A, Brouwers M, Charles C, Baldassarre F, Gauld M. Impact of cancer-related decision aids. Rockville, MD, USA: Agency for Healthcare Research and Quality. Evidence Report/Technology Assessment; 46. 2002 Indexing Status Subject indexing assigned by CRD MeSH Decision Making; Decision Support Techniques; Health Knowledge, Attitudes, Practice; Neoplasms; Patient Education as Topic; Patient Participation; Physician-Patient Relations AccessionNumber 12003008336 Date bibliographic record published 31/10/2004 Date abstract record published 31/10/2004 Record Status This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn. |
|
|
|